Genetica (2014) 142:323–336 DOI 10.1007/s10709-014-9777-2 Genetic structure in the Amazonian catfish Brachyplatystoma rousseauxii: influence of life history strategies F. M. Carvajal-Vallejos • F. Duponchelle • E. Desmarais • F. Cerqueira • S. Querouil • J. Nun˜ez • C. Garcı´a • J.-F. Renno Received: 21 December 2013 / Accepted: 7 July 2014 / Published online: 20 July 2014 Ó Springer International Publishing Switzerland 2014 Abstract The Dorado or Plateado (Gilded catfish) Bayesian analyses revealed at least three clusters in Brachyplatystoma rousseauxii (Pimelodidae, Siluriformes) admixture in the five locations sampled in the Bolivian is a commercially valuable migratory catfish performing Amazon, whereas only two of these clusters were observed the largest migration in freshwaters: from the Amazonian in the Western Amazon. Considering the migratory headwaters in the Andean foothills (breeding area) to the behaviour of B. rousseauxii, different life history strategies, Amazon estuary (nursery area). In spite of its importance to including homing, are proposed to explain the cluster dis- inform management and conservation efforts, the genetic tribution. Our results are discussed in the light of the variability of this species has only recently begun to be numerous threats to the species survival in the Madera studied. The aim of the present work was to determine the basin, in particular dam and reservoir construction. population genetic structure of B. rousseauxii in two regions: the Upper Madera Basin (five locations in the Keywords Bolivia Á Madera River Á Population Bolivian Amazon) and the Western Amazon Basin (one structure Á Microsatellites Á Siluriformes Á Pimelodidae regional sample from the Uyucalı´–Napo–Maran˜on–Ama- zon basin, Peru). Length polymorphism at nine microsat- ellite loci (284 individuals) was used to determine genetic Introduction variability and to identify the most probable panmictic units (using a Bayesian approach), after a significant Brachyplastystoma rousseauxii (Castelnau 1855) is a large departure from Hardy–Weinberg equilibrium was observed ([1.5 m) migratory catfish of the family Pimelodidae, in the overall dataset (Western Amazon ? Upper Madera). commonly known as Dorado in Bolivia, Colombia and F. M. Carvajal-Vallejos Á F. Duponchelle Á E. Desmarais Á F. Duponchelle Á S. Querouil Á J. Nun˜ez Á J.-F. Renno S. Querouil Á J. Nun˜ez Á C. Garcı´a Á J.-F. Renno Institut de Recherche pour le De´veloppement (IRD), UMR- Laboratoire Mixte International – Evolution et Domestication de BOREA, 911 Avenue Agropolis, 34394 Montpellier, France l’Ichthyofaune Amazonienne (LMI – EDIA), Centro de Investigaciones Quistococha, Iquitos-Loreto, Peru´ E. Desmarais Á F. Cerqueira Centre National de la Recherche Scientifique (CNRS), UMR- F. M. Carvajal-Vallejos ISEM (Institut des Sciences de l’Evolution de Montpellier), FAUNAGUA (Institute for Applied Research on Aquatic Universite´ Montpellier 2, Baˆtiment 22, C.C. 065, Resources), final Av. Max Ferna´ndez s/n, zona Arocagua Norte, 34095 Montpellier Cedex 5, France Municipio de Sacaba, Cochabamba, Plurinational State of Bolivia C. Garcı´a Instituto de Investigaciones de la Amazonı´a peruana (IIAP), F. M. Carvajal-Vallejos (&) Carretera Iquitos-Nauta, km 4.5, Quistococha, Iquitos, Peru´ Unidad de Limnologı´a y Recursos Acua´ticos (ULRA), Facultad de Ciencias y Tecnologı´a (FCyT), Universidad Mayor de San Simo´n (UMSS), calle Sucre frente al Parque La Torre s/n, zona Las Cuadras, Cochabamba, Plurinational State of Bolivia e-mail: [email protected] 123 324 Genetica (2014) 142:323–336 Peru, as Dourada in Brazil and as Plateado in Bolivia explanations since the sampling design and molecular (Carvajal-Vallejos et al. 2011). It is one of the most markers used (the maternally-inherited Control Region— emblematic species of the Amazon basin, owing to its CR at mtDNA) did not provide sufficient information or economic importance and exceptional life cycle, involving consistent evidence of a philopatric migration model (see the largest known migration in freshwater (Barthem and Campos Telles et al. 2011) and were unable to define the Goulding 1997, 2007; Alonso 2002; Araujo-Lima and reproductive units related to a possible homing behaviour. Ruffino 2003). The Dorado is widely distributed through- More recently, Batista (2010), using nuclear (microsatel- out the white-water tributaries and main channels of the lites) and mitochondrial (CR) DNA markers, concluded Amazon River (Araujo-Lima and Ruffino 2003; Barthem without testing departure from Hardy–Weinberg equilib- and Goulding 2007), and also can be observed sporadically rium (HWE) that the large number of individuals collected in clear- and black-water rivers (e.g. Negro, Tocantins— from different tributaries of the Amazon basin (including Goulding 1980). There are few studies on this species to the Madera River) were parts of a single panmictic unit. date, despite the interest that arose among ichthyologists as The possibility of a homing phenomenon, however, was a result of Barthem and Goulding’s (1997) hypothesis not totally dismissed, as some degree of structure was about its life cycle. These authors followed the size dis- observed for mtDNA among some large tributaries. These tribution of the Dorado among the main fishing ports along conclusions, suggesting the existence of a single population the Amazon and Lower Madera River channels, from the of Dorado in the Amazon basin, could have dramatic estuary (Belem) all the way to the headwaters of an consequences for species conservation. Indeed, the Dorado Amazon tributary in Peru (Pucallpa, 4,500 km from the is considered overexploited in the Lower Amazon (Petrere estuary). Fishery statistics and experimental fishing data et al. 2004; Alonso and Pirker 2005) and heavily exploited collected over several years showed a clear change in size in some areas of the Upper Amazon in Peru (Garcı´a et al. distribution of the Dorado between the estuary and the 2009a, b) and Colombia (Agudelo et al. 2013). If this headwaters, with only small and medium-sized fish in the species is composed of a large, single panmictic population estuary, almost exclusively large specimens in the head- of homogeneous individuals throughout its distribution waters of Peru or Colombia, and intermediate size-classes area, then over-exploiting the species in some tributaries in between. They also observed that juvenile Dorado were would not be particularly problematic at regional scale, as absent and pre-adults and adults were rarely found in the recruitment would be ensured from other, less exploited flood plains. Alonso (2002), using age determination from tributaries. On the other hand, if there is genetic structure otolith readings, reconstructed the age-size distribution of within the species, then the total allowable catches have to the species along the main axis of the Amazon River. His be set and enforced for each population. Additionally, the results confirmed that while mostly juveniles were present growing number of hydroelectric power plants under con- in the Amazon estuary, almost exclusively adults were struction in the Amazon basin are further threatening the observed in the Iquitos area and mostly pre-adults were species by disturbing or interrupting its migration routes caught in between. Providing new information regarding (Junk et al. 2007; Van Damme et al. 2011b; Finer and age, growth, reproductive biology, and species’ depen- Jenkins 2012). But again, if the single panmictic population dence on fluvial dynamics along the Amazon main stem, hypothesis is verified, a depletion of the breeding indi- studies of Garcı´a et al. (2009a) and Agudelo et al. (2013) viduals in the Upper Madera following the Madera confirmed the existence of a complex life cycle involving hydroelectric impoundments of Santo Antoˆnio and Jirau, long-distance migration between the Amazon estuary and might be regarded by dam supporters as not so dramatic the headwaters. In the Peruvian and Colombian Amazonia, (even though reducing the range of a species could be where these studies were carried out, B. rousseauxii problematic), as recruitment for the population would be reaches sexual maturity at about 4 years old and 90 cm SL ensured from the Upper Amazon–Solimo˜es system. Con- and has longevity of at least 15 years (Garcı´a et al. 2009; sidering that the Dorado is an economically (e.g. Goulding Agudelo et al. 2013). 1979; Agudelo et al. 2000) and ecologically (Angelini et al. Batista and Alves-Gomez (2006) suggested that the 2006) important, yet relatively poorly studied resource with Dorado displays spatial genetic variability along the a complex life cycle occurring over a continental scale Amazon’s mainstem that could be explained by a homing while subject to differential fishing pressures (on juveniles behaviour. These authors observed that haplotypic diver- in the Central Amazon Basin and on adults in the Western sity decreases from the estuary westwards and suggested Amazon Basin), it is crucial to better understand its biology that this could be due to preferential recruitment of dif- and genetic structure and shed new light on a potential ferentiated populations to certain tributaries as the fish homing behaviour. move upstream. While an interesting hypothesis, the vari- The Amazon Basin is over 6,000,000 km2, and its main ability observed in this study also could have alternative tributaries (e.g. Madera, Ucuyalı´) have distinct hydrological 123 Genetica (2014) 142:323–336 325 Fig. 1 Composition and distribution of the